The product changelog for CompetLab — competitive intelligence for B2B SaaS across 14 dimensions (6 monitored, 8 leading-edge), including AI Visibility: how ChatGPT, Claude, Gemini, Perplexity and Google AI Overviews name and recommend you against competitors. Each entry below is a shipped capability, newest first.
Two skills read the two checks that went live on 17 Oct: competlab-pricing, what each company charges as its page shows it, typical setups cheapest first, what changed; competlab-positioning, what each homepage says in its own words, what you say that nobody else does, what changed, what pages say about you.
npx skills add competlab/competlab-ci-skills --allStrategic Tickets puts the recommendations of CompetLab's Strategic Briefing on a board: new ones open as tickets in Triage, most important first, each with an impact rating, an effort estimate and what it is based on. The team decides which to take and moves them across five columns: Triage, To do, In progress, Done, Dismissed. From a project's second briefing on, the briefing reads the board before it writes. Where it finds a ticket already there for a recommendation, in any column, it opens no second one. It comments on a ticket only when CompetLab's monitoring has measured something new about it — what was measured and when, never what caused it. The board can be read and written over the REST API and MCP.
competlab.com/strategic-briefing-example?view=ticketsThe open-source competlab-ci-skills repository goes from thirteen skills to seven. The seven leading-indicator skills are retired: the Strategic Briefing now researches those areas itself, over fourteen analysis areas with history. The CMO Report orchestrator, the weekly briefing and the landscape skill become one Briefing skill that reads it. What remains reads the platform and shapes its output — AI Visibility, AI Sources (new), Briefing, Competitor Dive, Battlecard, Site Audit (new: any public domain, no project needed, a free-trial key is enough) and Monitoring Setup (new). Every skill ships the platform's reporting rules: null is not zero, figures as counts and a share only beside its range, per engine never pooled, retrieved never cited.
npx skills add competlab/competlab-ci-skills --allAI Sources fetches the pages Perplexity and Google AI Overviews retrieved while answering your buyers' questions, and reads them — reporting per engine which companies each engine named, and which of those pages name other companies and not you. Being retrieved and being named are two different events, and this is the dimension that tells you which one you are missing.
app.competlab.com/registerAI Visibility now asks five engines — ChatGPT, Claude, Gemini, Perplexity and Google AI Overviews — who they name, who they recommend, and how warmly. Google AI Overviews reaches buyers who never open a chat assistant, and CompetLab reads it the way a searcher sees it: the companies Google's own text names, in the order that text names them.
AI Visibility now runs on how often each engine recommends you — the share of answers naming you, read across a rolling window and published as a range rather than one number. A lead counts only when two ranges sit apart; where they overlap, CompetLab says the match is too close to call instead of naming a winner. The average position it replaces was retired on evidence rather than instinct: pooling the engines into one number produced movement that none of the engines actually had.
The AI Crawler Checker now answers which named AI assistants can reach a site — ChatGPT, Claude, Perplexity, Microsoft Copilot, Google AI Overviews and Gemini Apps — naming the crawler that decided each verdict and the line number in the site's own robots.txt behind it. The 0-100 score is gone: any weighting marks down sites that blocked training crawlers deliberately, which costs them nothing in whether assistants can cite them. Two crawlers break the usual training-versus-retrieval split and are reported as themselves: bingbot decides Microsoft Copilot, which has no crawler of its own, and Google-Extended fetches nothing at all while governing both model training and Gemini Apps. When a file cannot be read, the tool says so and withholds the verdict rather than reporting a failed request as an open site.
competlab.com/tools/ai-crawler-checkerFree tool that checks any site's llms.txt against the spec — or generates a clean one from its sitemap — plus the honest, data-backed truth about who actually reads it (and who doesn't).
competlab.com/tools/llms-txt-generator-checkerCompetLab's analyst layer: a Strategic Briefing that synthesizes 13 dimensions — the 5 it monitors continuously plus 8 it researches for the briefing — into one narrative: what changed, what it means, what to do next, and what to watch. It's generated for you the moment you create a project, then refreshed every month, readable in-app and downloadable as PDF or Markdown.
app.competlab.com/registerThe open-source competlab-ci-skills repository grows to 13 skills. Alongside the 5 core CI workflows (AI-visibility reports, briefings, battlecards, competitor dossiers, and landscape analysis), v2.0 adds 7 leading-indicator watches — funding, hiring, product velocity, agent-adoption, customer voice, status, and AI-ecosystem signals — plus an orchestrator that sequences all 13 into a single CMO-grade briefing.
npx skills add competlab/competlab-ci-skills --allFree tool that scores any domain on AI-agent adoption — MCP server presence, llms.txt, AGENTS.md, Schema.org, and agent-readable content.
competlab.com/tools/agent-adoption-checkFree tool that maps any site's sitemap into a visual content tree — structure, categories, and gaps at a glance.
competlab.com/tools/sitemap-visualizerFree tool that reads robots.txt to show which AI crawlers (GPTBot, ClaudeBot, Google-Extended, and more) a site allows or blocks.
competlab.com/tools/ai-crawler-checkerFree tool that surfaces a domain's security headers (graded A–F), trust markers, and credibility signals.
competlab.com/tools/trust-signals-analyzerFree tool that fingerprints any company's tech stack — frameworks, analytics, hosting, and infrastructure.
competlab.com/tools/tech-stack-analyzerPre-built competitive-intelligence workflows for AI coding agents (Claude Code, Cursor, Codex, Gemini CLI), powered by CompetLab's live data. MIT-licensed.
github.com/competlab/competlab-ci-skillsCompetLab's Model Context Protocol server exposes 38 tools so any AI agent can pull competitive intelligence directly.
mcp.competlab.com/mcpOfficial SDK on npm — 25 methods over the CompetLab API, fully typed for Node.js.
npm install @competlab/sdkProgrammatic access to every dimension of competitive data, with managed API keys.
competlab.com/developers/apiSelf-serve onboarding — start free, no credit card, with full subscription and billing built in.
app.competlab.com/registerCompetitive changes delivered instantly or as scheduled digests, on the channels your team already uses.
Automatic change detection across all 5 monitored dimensions, with severity scoring and exact diffs of what each competitor changed.
Tracks competitor homepage messaging, value props, CTAs, target audiences, and content/sitemap structure over time.
Monitors competitor plans, billing models, free tiers, and enterprise pricing — with side-by-side gap analysis.
CompetLab re-checks every competitor across all 5 monitored dimensions on a recurring schedule — automatically.
CompetLab's flagship dimension: track how ChatGPT, Claude, and Gemini mention and rank your brand against competitors over time, scored 0–100 (AI Visibility Score). The dimension CompetLab was built around.
Competitive intelligence for B2B SaaS — track competitors across 5 monitored dimensions from one automated dashboard.